How Free is Parameter-Free Stochastic Optimization?

Authors: Amit Attia, Tomer Koren

ICML 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical We study the problem of parameter-free stochastic optimization, inquiring whether, and under what conditions, do fully parameter-free methods exist: these are methods that achieve convergence rates competitive with optimally tuned methods, without requiring significant knowledge of the true problem parameters. [...] In the non-convex setting, we demonstrate that a simple hyperparameter search technique results in a fully parameter-free method that outperforms more sophisticated state-of-the-art algorithms. We also provide a similar result in the convex setting with access to noisy function values under mild noise assumptions. Finally, assuming only access to stochastic gradients, we establish a lower bound that renders fully parameter-free stochastic convex optimization infeasible, and provide a method which is (partially) parameter-free up to the limit indicated by our lower bound.
Researcher Affiliation Collaboration 1Blavatnik School of Computer Science, Tel Aviv University 2Google Research Tel Aviv.
Pseudocode Yes Algorithm 1: Adaptive projected SGD tuning; Algorithm 2: Non-convex SGD tuning; Algorithm 3: Convex SGD tuning
Open Source Code No The paper does not contain any explicit statement about open-source code availability for the described methodology, nor does it provide a link to a code repository.
Open Datasets No This is a theoretical paper focused on algorithm design and analysis, and as such, it does not use or provide information about specific datasets for training.
Dataset Splits No This is a theoretical paper focused on algorithm design and analysis, and therefore does not discuss training, validation, or test splits of datasets.
Hardware Specification No This is a theoretical paper that focuses on mathematical analysis and algorithm design, and as such, it does not specify any hardware used for experiments.
Software Dependencies No This is a theoretical paper focused on mathematical analysis and algorithm design; therefore, it does not list specific software dependencies with version numbers.
Experiment Setup No This is a theoretical paper focused on algorithm design and analysis, and does not provide details of an experimental setup such as hyperparameters or training configurations.